Identifying images using a tree structure

1. A method of identifying an image characteristic, comprising the steps of: storing a tree structured code having coding representing each of a plurality of training images; comparing coding representing a sample image, not included in the plurality of training images, to the tree structured code to identify the training image coding in the structured tree code closest to the sample image coding; and identifying a characteristic of the sample image based upon the training image represented by the closest training image coding.

2. A method according to claim 1 , wherein the characteristic is a pose.

3. A method according to claim 1 , wherein the plurality of training images include images having at least one of different subjects and different lighting conditions.

4. A method according to claim 1 , further comprising the step of: determining if an alarm is to be sounded based upon the identified characteristic.

5. A method according to claim 1 , wherein the plurality of training images include images of a plurality of different individual subjects of a particular type and the sample image includes an image of a subject of the particular type not included in the plurality of different subjects.

6. A method according to claim 1 , wherein the training image coding representing each of the plurality of training images in the tree structured code is n dimensional coding, where n is the number of pixels in that image.

7. A method according to claim 1 , wherein the n dimensional coding includes a vector quantization value corresponding to a vector representing the applicable image in n dimensional space.

8. A method according to claim 6 , wherein the sample image coding is n dimensional coding.

9. A method according to claim 8 , wherein the n dimensional coding includes a vector quantization value corresponding to a vector representing one of (i) the applicable training image in n dimensional space and (ii) the sample image, in n dimensional space.

10. A method according to claim 1 , further comprising the step of: separating the plurality of training images; wherein the coding representing each of the plurality of training images within the tree structured code is related to the other coding representing other of the plurality of training images within the tree structured code based on the separation of the plurality of training images.

11. A method according to claim 1 , further comprising the steps of: determining a direction of maximum variation for the plurality of training images; and determining a distance from the direction of maximum variation for each of the plurality of training images; wherein the coding representing each of the plurality of training images within the tree structured code is related to the other coding representing other of the plurality of training images within the tree structured code based on the determined distance for the training image represented by that coding.

12. A method according to claim 1 , further comprising the steps of: determining a direction of maximum variation for the plurality of training images; and determining a direction for each of the plurality of training images with respect to the direction of maximum variation; wherein the coding representing each of the plurality of training images within the tree structured code is related to the other coding representing other of the plurality of training images within the tree structured code based on the determined direction for the training image represented by that coding.

13. A method according to claim 1 , further comprising the step of: locating a window within the sample image, the window being located to have a subject, depicted in the sample image contained within the window; wherein the coding representing the sample image represents only a portion of the sample image within the window.

14. A method according to claim 13 , wherein the window is located so as to center the subject within the window.

15. A method according to claim 13 , wherein the sample image is one of a plurality of sample images, each of the plurality of sample images having the subject depicted in that sample image, and further comprising the step of: locating the window within each of the plurality of sample images to have the subject depicted in that sample image contained within the window.

16. A method according to claim 15 , wherein the window is located in each of the plurality of sample images so as to have the subject contained at approximately a same position within the window located within each of the sample images.

17. A method according to claim 15 , wherein: the plurality of training images include at least one training subject depicted at different locations in the plurality of training images; the window is located within at least one of the plurality of sample images based on the at least one training subject being depicted at a particular one of the different locations in the plurality of training images.

18. A method according to claim 1 , wherein the plurality of training images include a training subject depicted in different positions in the plurality of training images, and the sample image is one of a plurality of sample images, each of the plurality of sample images having a subject depicted in that sample image, and further comprising the steps of: positioning a window within a first of the plurality of sample images, wherein with the window so positioned the subject in the first sample image is at a first position within the window which is different than a predetermined position; and positioning the window within a second of the plurality of sample images, the window being positioned within the second sample image differently than positioned within the first sample image, the difference in the positioning corresponding to the difference between the first position and the predetermined position, wherein with the window positioned within the second sample image the subject in the second sample image is at the predetermined position within the window; wherein the coding representing the second sample image represents only a portion of that sample image within the window.

19. A method according to claim 18 wherein: the plurality of training images include a training subject depicted in different positions within the plurality of training images; and the window is positioned within the second sample image based on the training subject being at a second position within the plurality of training images which corresponds to the first position.

20. A method according to claim 18 , wherein the window is positioned differently within the second sample image with respect to one of (i) window location, (ii) window orientation, and (iii) window size.

21. A system for identifying an image characteristic, comprising: a memory configured to store a tree structured code having coding representing each of a plurality of training images; and a processor configured to process coding representing a sample image, not included in the plurality of training images, to identify a characteristic of the sample image by traversing the tree structured code to locate the training image coding in the structured tree code which is closest to the sample image coding.

22. A system according to claim 21 , wherein: the processor is further configured to position a window within the sample image so as to have a subject depicted in the sample image disposed within the window; and the coding representing the sample image represents only that portion of the sample image within the window.

23. A system according to claim 22 , wherein: the sample image is one of a plurality of sample images, each of the plurality of sample images having the subject depicted in that sample image; and the processor is further configured to position the window within each of the plurality of sample images to have the subject depicted in that sample image disposed within the window.

24. A system according to claim 23 , wherein: the processor is further configured to position the window within each of the plurality of sample images so as to have the subject disposed at approximately a same position within the window.

25. A system according to claim 23 , wherein: the plurality of training images include at least one training subject disposed at different positions in the plurality of training images; and the window is located within at least one of the plurality of sample images based on the training subject being disposed at a particular one of the different positions within the plurality of training images.

26. A system according to claim 21 , wherein: the plurality of training images include a training subject disposed at different positions in the plurality of training images; the sample image is one of a plurality of sample images, each of the plurality of sample images having a sample subject depicted in that sample image; the processor is further configured (i) to position a window within a first of the plurality of sample images such that the sample subject in the first sample image is at a first position within the window which is different than a predefined position, and (ii) to position the window within a second of the plurality of sample images so as to be positioned with the sample subject in the second sample image at the predetermined position, based upon the difference between the first position and the predetermined position; and the memory is further configured to store only the coding representing that portion of the each of the plurality of sample images within the window.

27. A system according to claim 26 , wherein: the plurality of training images include a training subject disposed at different positions in the plurality of training images; and the processor is further configured to position the window within the second sample image based on the training subject being at a second position, corresponding to the predefined position, within the plurality of training images.